New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Are Medusa Heads computed in parallel or serially? #116

Open

userljz opened this issue Aug 3, 2024 · 0 comments

userljz commented Aug 3, 2024 •

edited

Loading

Hello authors,

While reading your code, I noticed that the multiple Medusa Heads you proposed are computing results in parallel

for i in range(self.medusa):
    medusa_logits.append(self.medusa_head[i](hidden_states))

(although the later Heads don't use the results from the previous Heads, the results are obtained using a for loop).

I'm wondering if I've misunderstood this, or if Medusa is currently using serially obtained results?

Could you please clarify this for me?

The text was updated successfully, but these errors were encountered:

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment